MATERIALS AND METHODS

throughout lactation period.

MATERIALS AND METHODS

Animals

The experiment was carried out from August 2009 to February 2010 at the Veterinary Medical Park and Animal Science Teaching and Research Center (Department of Animal Science, Texas A&M University) using 18 multiparous sows (Landrace X Large White) and their litters. Three days prior to the expected day of parturition, sows were transferred into individual farrowing pens. The experimental protocol of this study was approved by the Texas A&M University Institutional Animal Care and Use Committee.

Housing and Management

Each sow was housed with her piglets in a farrowing pen (1.8 m X 2.5 m = 4.5 m² including 1.98 m² of plastic-coated perforated floor). Each pen had metal sidewalls and was equipped with a sow feeder (0.45 m X 0.35 m) and a nipple drinker. During their lactation, sows were fed up to 7.3 kg d, separated into two feeding periods (08:00 am and 17:00 pm).

The diet contained 3.30 MCal ME/kg and 18.7% protein which met the NRC requirement for lactating sows (Mateo et al., 2008). Sows had ad libitum access to water during the 35 d lactation period. Ambient air temperature ranged from 22-32℃; radiant heat was provided to the litter as needed.

At birth, the number and gender of piglets born both alive and dead were recorded.

Mean litter size was 8.9±1.0 piglets. Teeth were clipped and tails were docked at d 3 of age.

Each piglet was also marked individually with ear notches and received an iron injection.

During the whole lactation period, piglets had free access to water. They were not given creep feed during the 35 d lactation period.

Data Collection and Statistical Analysis

Eighteen litters were involved in the study. All piglets (n=160) were weighed immediately after birth and on day 7, 14, 21, 28 and 35 of the lactation period. Average daily gain was calculated. On the day of weighing, three consecutive periods were analyzed from 08:00 to 17:00. Piglets were isolated from the sow for 2 hr then nursed for 1 hr and this procedure repeated three times. Milk consumption of the piglets was estimated using weight-suckle-weight (Wu G. et al., 2000). Piglets were classified into four categories according to their birth weight. A: 0.7-1.09 kg; B: 1.10-1.49 kg; C: 1.50-1.89 kg; and D: >1.90 kg. Piglets with a birth weight of 0.7 – 1.09 kg are classified as intrauterine growth retardation (IUGR) (Wu et al., 2010). Six analyses were carried out on different measures for these 4 groups. All the analyses are based on One-Way ANOVA and the Duncan‟s multiple range tests. ADG, body weight and milk intake of piglets were analyzed with birth weight as a fixed effect on the given days. When the piglets died, their body weight and the time of death were recorded to calculate mortality. The data from dead piglets then were excluded from further analyses.

RESULTS

Piglets Body Weight

Body weights of each of the piglets were measured at birth, with the time at birth denoted as D0. The mean birth weight for each category was calculated. Afterwards, all weaning piglets were weighted every 7 days over 35 days of lactation. For D0, D7, D14, D21, D28 and D35, ANOVA indicated that the means of the four categories on each day differed significantly (P<0.01): mean D > mean C > mean B > mean A (Table 1). Mortality was highest in the lowest birth weight group especially in the first week of study (Table 2).

For the first week of study, mortality rate was 26.1 percent for the lightest birth weight group vs 4.26, 1.69 and 0 percent for other groups respectively. An exponential function was used to model the constant change in the independent variable (days) with the proportional change (increase) in the dependent variable (body weight). The graph of y = e^x was upward-sloping, and increased faster as x increased. The body weight had an exponential relation to the growth; therefore, the results showed that for the heaviest birth weight group, body weight increased in a decreasing way. The body weights of lighter pigs increased relatively faster but were still significantly the smallest during the 35 d of lactation (Figure 1).

35

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

*

D 0 = Day of birth.

†

Number of piglets at birth.

a-d

Means not sharing a common superscript within a column differ (P<0.01).

Table 2. Numbers of piglets at each week interval^*

35 Figure 1. Body weight of piglets during the study

35 Percentage Increase in Body Weight

Relative growth rate of each piglet in each week was calculated based on the difference of body weight between last and first day of the week. The time intervals are denoted as D (0-7), D (7-14), D (14-21), D (21-28) and D (28-35) indicating 7 days intervals between each measurement. All ANOVA tests, each of which is for one of these five time slots, show that at least the means of two of the categories are different from each other (P <

0.05). For D (0-7), mean B > mean A > mean C > mean D (P<0.01). For D (7-14), the result is mean A = mean B = mean C > mean D with significance. And for D (14-21), D (21-28) and D (28-35), mean A > mean B > mean C > mean D (Table 3). These results indicate that the growth rate decreases at a rate proportional to initial period body weight (Figure 2).

According to these results, the percentage of growth of lighter birth weight piglets was higher than their heavier contemporaries in each week if they could successfully survive during the first week of neonatal period.

Table 3. Percent increase in body weight of piglets

Groups of piglets n^† D(0-7) D(7-14) D(14-21) D(21-28) D(28-35)

Changes in BW (%/week)

A: 0.7-1.09 kg 23 71.4±6.6^b 54.7±7.2^a 41.5±2.9^a 28.1±1.5^a 21.1±0.9^a B: 1.10-1.49 kg 47 74.9±4.2^a 53.5±2.0^a 37.3±2.2^b 26.4±0.9^b 19.9±0.5^b C: 1.50-1.89 kg 59 60.4±2.5^c 54.0±1.6^a 33.8±1.2^c 24.6±0.8^c 19.5±0.4^c D: >1.9 kg 31 54.8±3.1^d 46.2±2.0^b 30.1±1.4^d 22.7±0.9^d 18.2±0.7^d

P value < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

†Number of animals at birth.

a-d Means not sharing a common superscript within a column differ (P<0.01).

35 Figure 2. Growth rate of piglets measured as percentage increase in body weight in each

week

Absolute Changes in Body Weight (unit: g/d)

Average daily gain (ADG) is a significant variable in assessing growth rate in animals. All ANOVA tests indicated that at least two of the means differed (P < 0.05) significantly for each one of the four initial weight categories. Based on the Duncan‟s tests, the result for D (0-7) is mean D > mean C > mean B > mean A. After week one mean D = mean C > mean B > mean A (Table 4) (P<0.01). Overall, the two heaviest groups were not different from each other in terms of absolute changes in body weight, but their magnitudes of changes in body weight were larger than the two comparatively lighter groups in a statistical context (Figure 3). The daily body weight gain (g/day) reached peak values for all groups between the second and third week of age and after that ADG for all groups started to decrease (Figure 4).

35 Table 4. Absolute changes in body weight or daily weight gain (ADG) of piglets

Groups of piglets n^† D(0-7) D(7-14) D(14-21) D(21-28) D(28-35)

Changes in BW (g/d)

A: 0.7-1.09 kg 23 102±10^d 141±19^c 168±15^c 163±14^c 157±14^c

B: 1.10-1.49 kg 47 136±8^c 169±7^b 182±11^b 181±9^b 173±8^b

C: 1.50-1.89 kg 59 145±6^b 209±8^a 199±7^a 195±7^a 193±6^a

D: >1.9 kg 31 163±9^a 212±10^a 200±9^a 197±8^a 194±8^a

P value < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

†Number of animals at birth.

a-d Means not sharing a common superscript within a column differ (P<0.01).

Figure 3. Comparison of average daily gain between groups

abcd mean not sharing a common superscript difference (P<0.01)

35 Figure 4.Average daily increase in body weight of piglets (g/day)

Milk Intake (unit: g/kg BW)

Using weight-suckle-weight technique, the average milk intake for the piglets in each category was recorded on days 7, 14, 21, 28 and 35. The milk intake then was calculated based on the total daily milk intake of the piglets and their body weight. All of the ANOVA tests indicated at least two of the means were different significantly for each one of the four categories. Based on the Duncan‟s test, for D7 and D14, mean D = mean C > mean B >

mean A significantly. For D21, mean D > mean C > mean B = mean A with significance. For D28 and D35, mean D > mean C > mean B > mean A with significance (Table 5) (P<0.01).

Overall, heavier piglets at birth consumed more milk per kg body weight. The highest rate of milk intake by piglets was at D7 followed by an exponential decrease over next three weeks (Figure 5). Thus, there was a limitation in milk production. As piglets grew older, their milk intake per kg body weight decreased dramatically.

35 Table 5. Milk consumption per kg of body weight

Groups of piglets n^† D 7 D 14 D 21 D 28 D 35

Milk Intake (g/kg BW)

A: 0.7-1.09 kg 23 270±11^c 187±8^c 148±4^c 120±3^d 101±3^d

B: 1.10-1.49 kg 47 276±7^b 196±4^b 150±3^c 125±3^c 106±3^c

C: 1.50-1.89 kg 59 282±6^a 210±3^a 161±3^b 132±2^b 112±2^b

D: >1.9 kg 31 284±11^a 211±7^a 166±6^a 138±5^a 119±4^a

P value < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

†Number of animals at birth.

a-d Means not sharing a common superscript within a column differ (P<0.01).

Figure 5. Milk consumption of the piglets per kg body weight for each group

Daily Milk Consumption (unit: g/d)

We recorded the average milk consumption which was measured as grams per day for individual piglets in each group on days 7, 14, 21, 28 and 35 (Table 6). All of the ANOVA

35 tests at least two of the means were different significantly for each one of the four groups.

Based on the Duncan‟s test, mean D > mean C > mean B > mean A (P<0.01) (Table 6). As piglets became older, they consumed more milk and this increasing trend was higher in heavier piglets during the 35 d study (Figure 6).

Table 6. Milk consumption of piglets g/day/pig

Groups of piglets n^† D 7 D 14 D 21 D 28 D 35

Milk Intake (g/d)

A: 0.7-1.09 kg 23 475±42^d 525±51^d 600±46^d 615±44^d 628±42^d

B: 1.10-1.49 kg 47 639±27^c 686±24^c 723±26^c 757±32^c 767±31^c

C: 1.50-1.89 kg 59 772±27^b 875±24^b 897±24^b 910±24^b 925±24^b

D: >1.9 kg 31 922±45^a 999±46^a 1021±45^a 1039±44^a 1055±44^a

P value < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

†Number of animals at birth.

a-d Means not sharing a common superscript within a column differ (P<0.01).

35 Figure 6. Daily consumption of milk by piglets

Milk Efficiency (Average daily weight gain/milk intake)

We measured the conversion of milk to body weight for the piglets in each category in five time intervals which are defined as D (0-7), D (7-14), D (14-21), D (21-28) and D (28-35). All ANOVA tests indicated that at least two of the means were different significantly for each category (P<0.01). The results of the Duncan‟s tests showed fluctuation for the first two time slots, but became quite neat and stable for the last three intervals. Specifically, for D (0-7), mean B = mean D > mean C > mean A with statistical significance. For D (7-14), mean B > mean C = mean A > mean D significantly. For the rest three time intervals, mean A > mean B > mean C > mean D with significance (Table 7).

44 Table 7. Efficiency of conversion of milk to body weight gain

Groups of piglets

n^† D(0-7) D(7-14) D(14-21) D(21-28) D(28-35)

Average daily gain/milk intake; g/g

A: 0.7-1.09 kg 23 0.184±0.011^c 0.167±0.028^b 0.192±0.011^a 0.178±0.009^a 0.167±0.009^a B: 1.10-1.49kg 47 0.212±0.008^a 0.173±0.007^a 0.168±0.008^b 0.161±0.006^b 0.151±0.005^b C: 1.50-1.89kg 59 0.206±0.006^b 0.167±0.004^b 0.149±0.004^c 0.144±0.004^c 0.140±0.003^c D: >1.9 kg 31 0.210±0.009^a 0.147±0.006^c 0.134±0.006^d 0.130±0.006^d 0.125±0.005^d

P value < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

Data are expressed as means ± SEM. The P value represents statistical significance among BW groups for each 7 d interval.

†

Number of animals at birth.

a-d

Means not sharing a common superscript within a column differ (P<0.01).

44 DISCUSSION

Body Weight and Growth

The present study offers a new perspective on the effect of birth weight on future economically important traits (e.g., mortality and growth rate) in that it provides an analysis of birth weight as a continuous effect. In general, our findings support the previous conclusion that low-birth-weight pigs have low absolute growth rate (g/day) during the suckling period (Gondret et al., 2005; Smith et al., 2007; Bérard et al., 2008; Rehfeldt et al., 2008). Similar to these studies, pigs were weaned in the current work at an older age (28 and 35 days) than typically practiced in commercial U.S. swine production (around 21 days).

Despite the use of different analyses, sample sizes, and facilities, light birth weight pigs were smaller at weaning compared to heavy birth weight pigs. Extending these reports, we noted that the relative growth rate (%/day) of surviving IUGR piglets was higher than their normal-weight counterparts (Figures 1 and 2).

Similar to the impact of birth weight on BW at weaning, regardless of weaning age, birth weight has also been shown to influence BW at various stages later in life. Dividing birth weight into categories, Powell and Aberle (1980) reported fewer days to 26 kg and 96 kg BW, Quiniou et al. (2002) reported increased BW at 63 d of age and fewer d to 105 kg, Smith et al. (2007) reported increased BW 42 d post weaning, and Rehfeldt et al. (2008) demonstrated heavier BW at 70, 133, and 180 (harvest) d of age due to greater birth weight.

However, in the present study, low-birth-weight pigs that successfully survived during preweaning showed a higher rate of growth as percentage increase in each week although they did not catch up in absolute body weight (Figure 2).

While not discussed by Rehfeldt et al. (2008), it appears there was a numeric disparity in the difference between heavy vs. middle categories compared to the difference between middle vs. light categories where the difference between light vs. middle birth weight was greater than the difference in middle vs. heavy category. This would agree with the present study findings of a greater difference in future BW for the lowest birth weight pigs (Table 1).

Again, using a cubic model, Schinckel et al. (2007) reported BW at 126 and 168 d of age, increased at a decreasing rate as birth weight increased, similar to our observation at day 35.

44 Regardless of the number of animals used, increased birth weight resulted in

increased BW measured at various stages during the study period (Figure 1). However, it is important to examine differences in future BW across the birth weight distribution and realize that an increase in birth weight from 0.7 to 1.0 kg does not have the same effect as an increase from 2.0 to 2.2 kg. From these results, it is apparent that birth weight affects future BW. The major cause of this appears to be a difference in ADG. Based on results from the current study, birth weight affects ADG during all phases of production; the difference is greatest for the lightest birth weight pigs (Figure 3). The findings of differences in ADG are in agreement with Rehfeldt et al. (2008). Not only do heavier pigs begin life with an advantage in weight but pigs at the lower end of the birth weight distribution, due to reduced ADG, fall further behind in BW over time (Table 3).

Several factors, both prenatal and postnatal, are likely responsible for this decrease in future growth due to reduced birth weight. First, low-birth-weight pigs had a smaller number of muscle fibers than heavier birth weight pigs (Nissen et al., 2004; Gondret et al., 2005, 2006;

Rehfeldt and Kuhn, 2006). The number of muscle fibers is determined prenatally; however, the increase in the size of the muscle fiber impacts growth (Dwyer et al., 1993; Rehfeldt et al., 2000; Herfort Pedersen et al., 2001). Additionally, rates of protein synthesis in skeletal muscle are lower in IUGR pigs than normal-birth-weight pigs during the suckling period (Junjun Wang and Guoyao Wu, unpublished data). Thus, because the growth of pigs is critically dependent on protein accretion in skeletal muscle, fewer muscle fibers in IUGR pigs would result in reduced future growth. Secondly, colostrum from the sow provides the newborn piglet with vital energy and maternal antibodies (Le Dividich et al., 2005). Several studies have reported at least a minor relationship between increased birth weight and the selection of anterior teats (McBride, 1963; Fraser, 1975; Hartsock et al., 1977) which have been shown to produce a greater amount of colostrum (Fraser and Lin, 1984) and pigs have reportedly gained more BW when nursing anterior teats (Kim et al., 2000). Other studies have measured colostrum intake and reported reduced intake due to low birth weight (Devillers et al., 2005, 2007). This difference in both colostrum and milk intake may contribute to the increase in pre-weaning ADG of pigs with heavier birth weight. The increase in ADG prior to weaning leads to heavier BW at weaning, which has been shown to result in increased post-weaning gain (Klindt, 2003). All of these factors contribute to the

44 reduced future BW and ADG due to reduced birth weight of piglets. However, the

relationship between birth weight and growth is not linear (Figure 1). There appears to be a threshold for birth weight where once surpassed, further increase in birth weight does not result in increased BW or ADG, especially later in life.

Pre-weaning Mortality

Increased birth weight was associated with a reduced (P < 0.01) chance of mortality prior to weaning. The greatest impact of birth weight on pre-weaning survival was for pigs with the lowest birth weights (Table 2). These findings are in agreement with other studies (Pettigrew et al., 1986; Gardner et al., 1989; Roehe and Kalm, 2000; Quiniou et al., 2002).

Increased pre-weaning mortality due to reduced birth weight could be attributable to a variety of prenatal developmental and postnatal environmental factors. Two postnatal factors which have been reported in the literature are vitality and food intake. De Roth and Downie (1976) reported lower birth weight pigs were given lower, or poorer, viability scores immediately following birth, and were more likely to suffer pre-weaning mortality. Lower birth weight pigs consume less colostrum and are more likely to suffer pre-weaning mortality (Devillers et al., 2005; Devillers et al., 2007). Other than crushing of piglets by the sow, early life mortality is presumably attributable to insufficient colostrum consumption (Le Dividich et al., 2005). While most pre-weaning mortality occurs early in lactation (Roehe and Kalm, 2000), there is also the potential for pigs to fall behind due to reduced milk consumption.

Lighter birth weight pigs have been shown to be at a competitive disadvantage and subject to less milk consumption (Hartsock and Graves, 1976). Milk intake is vital to the survival of the piglet. Our results also confirmed heavier birth weight pigs consume significantly more milk (P<0.01) than the lighter piglets throughout study (Table 5 and Table 6). Despite the availability of creep feed, pigs are mostly dependent on the sow to meet their nutrient and energy requirements prior to weaning (Sørensen et al., 1998).

Many studies have shown increased pre-weaning mortality in older sows (Gardner et al., 1989; Roehe and Kalm, 2000; Knol et al., 2002). This could be attributable to reduced litter size on younger sows; number of fully formed pigs could be affected by parity, which may account for a portion of the variation associated with litter size. Another explanation

44 may be the increased farrowing duration of older parity sows described previously (Canario

et al., 2006). The extended parturition of older sows could also lead to weaker pigs that survive through birth but are compromised and consequently susceptible to pre-weaning mortality. Roehe and Kalm (2000) and Rydhmer et al. (2008) showed preweaning mortality increases when gestation period decreases. Reduced gestation interval may result in less physiologically mature pigs being born and are more susceptible to pre-weaning mortality.

Colostrum production has also been associated with sow age and induced litters with short gestation lengths. The reduction in colostrum production led to reduced consumption by the piglets and adversely affected pre-weaning mortality (Devillers et al., 2005; Devillers et al., 2007).

To our knowledge, no result has been previously reported regarding milk conversion ratio or the relative rate of growth in each week for low birth weight piglets that successfully survived during preweaning period. Our data showed that surviving low-birth-weight piglets are, to a large extent, more efficient in milk utilization after day 14 (Table 7) and grow relatively faster in each week (Table 3) compared heavier birth-weight piglets. One possible explanation of this novel observation could be the lower maintenance requirement of energy and protein for low-birth-weight piglets due to reduced mass of the small intestine and liver, as well as reduced rates of protein degradation and amino acid oxidation in the whole-body, when compared to their bigger littermates. Also, bigger piglets require more energy and amino acids for their maximal performance during late lactation when production of sow‟s milk is limited. Given the advantage of surviving IUGR piglets in utilizing milk more efficiently for BW gain during the suckling period, every effort must be made to reduce neonatal mortality particularly in the first week of postnatal life.

In conclusion, our results confirms that low birth weight piglets grow in a slower rate and have a higher preweaning mortality compared to heavier birth weight piglets. Notably, IUGR piglets that survived successfully during the preweaning period can utilize milk more efficiently than bigger ones. The findings of the present work are significant in that they provide a new database for future studies to elucidate the biochemical mechanisms

In conclusion, our results confirms that low birth weight piglets grow in a slower rate and have a higher preweaning mortality compared to heavier birth weight piglets. Notably, IUGR piglets that survived successfully during the preweaning period can utilize milk more efficiently than bigger ones. The findings of the present work are significant in that they provide a new database for future studies to elucidate the biochemical mechanisms